Replies: 2 comments 8 replies
|
I am open to changing things in case it helps. Have you tried to play around with interning (does it actually help over gzip compression?) and protobuf to see how much it reduces trace sizes? I am little suprised to see such large profiles, I rarely see anything above 50MB. How many actions do you have in these large builds? |
8 replies
|
I understand your main concern is storage, but i have built ztracing to load and render large profiles which you might be interested in. We are embedding this viewer internally in order to handle large profiles (a few GB after decompression). I'm also open to change the format. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
On our end, we started seeing some larger users with Bazel profiles hitting 200 MB in the gzipped format.
I think this might be a good time to start a discussion about ways to improve the profile format for storage and transfer.
I think an easy win while keeping the existing trace viewer compatibility would be adopting Perfetto protobuf format and best practices https://perfetto.dev/docs/reference/synthetic-track-event#interning-data-for-trace-size-optimization. But I'm open to other ideas as well.
All reactions